1. Field of the Invention
This invention relates generally to electronic devices and the packaging thereof, including but not limited to hermetic and/or biocompatible packaging or sealing that renders such electronic devices environmentally isolated and/or implantable in vivo or in vitro. More particularly, the invention relates to electronic devices that include flexible circuit structures integrally bonded to and embedded with die structures, and assemblies or systems that include or cooperate with such electronic devices.
2. Related Art
An ongoing need exists for advances in electronic device miniaturization, packaging of integrated circuits and associated microfabrication technologies to enable new and improved implementations of electronic devices. Up to the present time, flip-chip technology as applied to bare dies has been considered to be the assembly process achieving the highest packing density, smallest footprint and lowest profile. For a detailed discussion of trends and challenges in state-of-the-art chip packaging (chip scale packaging, CSP and flip-chip), see Fjelstad et al., Chip scale packaging for modern electronics, Electrochemical Publications Ltd. (2002); Quinones et al., Flip-chip and chip scale packaging technologies: A historical perspective and future challenge, SEMICON China 2000 Technical Symposium: A1-A9 (2000). Briefly, in the flip-chip process the die is assembled face down onto the substrate (rigid board or flexible) with an array of solder bumps making electrical connection to the substrate. Reliability of the interconnects in flip-chip has been a great concern in the industry. This is particularly true for bonding to organic substrates because of the difference in thermal expansion coefficients between the chip and the substrate. Usually epoxies are used as underfill materials to make structures mechanically stronger and to improve reliability.
The need for further advances extends to environmentally isolated and/or biocompatible devices. Examples of biocompatible devices include implantable neural prostheses designed to interface with the nervous system to restore lost functions such as movement, hearing or vision. Other examples of neurostimulation devices include, without limitation, devices envisioned for spinal cord stimulation, deep brain and vagus nerve stimulation, sacral nerve stimulation, and gastric electrical stimulation. The requirements for such advanced implants are very different from implant devices found in the market today. The best known and commercially most successful implant is the cardiac pacemaker developed more than 30 years ago. Like the pacemaker, many of the FDA-approved implant devices use rigid packaging like titanium or ceramic casings for hermetic sealing and are equipped with mostly single or low-density microelectrodes for sensing and delivery of electrical stimulation. The electrodes and insulation are “oversized” and the devices are engineered for minimal failure incidents. Therefore, such devices are very bulky, limited in their functionality and require very invasive implantation procedures.
To enable new neurotechnology devices to interface effectively with the nerve system and to open up new applications, miniaturized and more flexible device structures with improved spatial and temporal sensitivity and packaging are needed. For many important applications such as the retinal implant where the shape of the implant structure needs to adapt to the curved shape of the inner eye, the substrate should be flexible to conform to the natural soft-tissue structure. That is, many types of implantable devices should ideally be biocompatible not only in the sense of chemical and biological inertness, but also in the sense of “mechanical” or “structural” biocompatibility, i.e., sufficient physical conformability and flexibility so as not to interact with surrounding tissue in an unwanted manner.
In recent years, researchers have focused on the use of flexible substrates like polyimides (or more recently liquid crystal polymers, or LCP) with hybrid assemblies of electronic chips, conduction layers and microelectrode arrays for stimulating and recording, all integrated on the flex substrate. A small footprint, high functionality, high reliability and biocompatibility are desirable attributes for active medical implants. The smaller the device, the less invasive is the procedure of implantation and one can expect better compatibility with surrounding tissue. At the same time, as noted above devices need to be packaged to have a biocompatible tissue interface and to withstand biodegradation in the body. The term biocompatibility in this context refers to mechanical biocompatibility as well as immunological biocompatibility.
The development of flexible polymer carriers for mounting and interconnecting chips and miniaturized components offer the possibility to develop micro-electronic and micro-optical systems that are in direct contact with delicate soft tissues and biological structures. Instead of standard housed IC components, bare or “naked” silicon chips and dies are used for hybrid integration to minimize component dimensions.
Unresolved critical issues in all implantable biomedical applications—and particularly for implanted neural prostheses and other devices envisioned to include flexible polymeric substrates—include packaging, integration and electrical connection of silicon dies or surface mounted electronics with the substrate and electrode arrays. Major challenges include bonding of chips or dies to flexible substrates and packaging of the device including electrical interconnects, connection pads and cables/leads. Protecting the implant from the corrosive effects of the biological fluids has been a particular challenge. Insulating biomaterials intended for implants need to protect devices from the hostile body environment for the lifetime of an implant recipient, sometimes for decades. Not only does the packaging need to withstand biodegradation in the body, but also the materials need to be biocompatible to prevent adverse reactions from the surrounding tissue. To date, the goal of a functional neural implant device that can survive for years in-vivo or in vitro has not been achieved.
Several companies use flip-chip processes to produce miniaturized turnkey electronic assemblies for medical implants, micro/miniature wireless devices, and a host of other applications. In addition to solder bumps, flip-chip technology has also employed gold wire stud bumps. Additionally, Parylene coatings have been employed for passivation to assemble a prototype implantable retina prosthesis with secondary receiving power and data coils. Researchers at the University of Utah are developing flip-chip assembly techniques to surface-mount chips directly on the back of a Si probe. See Solzbacher F., Chronic microelectrode arrays, Contract NINDS-NIH N01-NS-4-2362 (2004-2008). As a modification of traditional flip-chip processing, a group at the Fraunhofer Institute for Biomedical Engineering, St. Ingbert, Germany has developed a flexible interconnection technology to interconnect chips and surface-mount passive devices (SMD) with ultra-thin highly flexible polyimide (PI) substrates for a retinal implant using gold balls instead of solder bumps to connect the IC and substrate. See Stieglitz et al., Micromachined, polyimide-based devices for flexible neural interfaces, Biomed. Microdevices. 2(4): 283-294 (2000); Meyer et al., High density interconnects and flexible hybrid assemblies for active biomedical implants, IEEE Trans. Adv. Pack. 24: 366-374 (2001). This new assembly process is known as MicroFlex Interconnection (MFI). First, the PI substrate with metal traces and connection pads with a central via is microfabricated. The vias are aligned with the bond pads of the IC and a gold ball is bonded through the vias in the PI onto the metal pads of the chip utilizing a common thermosonic ball-bumping process. The gold ball acts as a stud or metal “rivet” to electrically connect and mechanically fix the chip or SMD to the substrate. This is a similar bonding scheme to flip-chip with gold studs replacing solder bumps. Because bonding occurs only at the through-via sites in the PI cable, an epoxy material is filled between the ribbon cable and the IC or SMD to improve stability of the connection. See Stieglitz et al., Micromachined, polyimide-based devices for flexible neural interfaces, Biomed. Microdevices. 2(4): 283-294 (2000).
Known technologies such as discussed above have not adequately addressed the above-mentioned problems. For instance, even with the use of underfill material, solder bumps, metal balls, rivets, and other conventional interconnects still represent potentially weak connection points, both structurally and electrically, between the chip or die and underlying substrate. These types of interconnects as well as the underfill material may still be prone to degradation in an environmentally or biologically hostile environment. Accordingly, mechanical stability, operational or functional reliability, biocompatibility, service life, etc. are still compromised in conventional packaged electronic devices. Moreover, sufficient miniaturization as needed for advanced devices such as electrostimulation devices has not been attained. As an example in the case of an intraocular implant such as an artificial retina, it is estimated that up to 1000 electrical neurostimulation sites are needed to restore useful vision in blind people. See Margalit et al., Retinal prosthesis for the blind, Survey Ophth. 47: 334-356 (2002). Currently, state-of-the-art retina chips have been designed and fabricated in the standard CMOS process with 1.5-μm feature size through the MOSIS foundry (Marina Del Rey, Calif.) to address up to 64 sites on a 4.6 mm×4.6 mm chip at the University of Michigan. See Ghovanloo et al., A modular 32-site wireless neural stimulation microsystem, IEEE J. Sol. State Cir. 39:1-10 (2004). Other groups in the USA and in Europe have built chips with similar capabilities and size. See Liu et al., A neuro-stimulus chip with telemetry unit for retinal prosthetic device, IEEE J. Sol. State Cir. 35: 1487-1497 (2000). It is not clear whether current technology is sufficient to enable engineers to design and build the circuitry to stimulate 1000 sites and package the circuitry into a single chip with this footprint because of limitations in the I/O and connections to substrate and the required voltage to stimulate neurons. Most likely, through the use of current technology, several chips would need to be mounted on the flex substrate to attain the required performance and functionality.
Therefore, in view of the foregoing, despite some advances in microfabrication technologies pertaining to packaged electronic devices, it is well-recognized by persons skilled in the art that an ongoing need exists for providing improved packaged electronic devices and related methods, apparatus and systems.